In [1]:
import pickle 
import matplotlib.pyplot as plt

from plots.finance import *
from plots.finance.logy_graph import LogYGraph, SeabornFn
from plots.finance.distribution_report import *
from plots.finance.risk_return_report import Risk_Return_Ratio_Report, TwoVariablePortfolioReport
from plots.finance.percentile_comparision import *
from plots.finance.efficient_frontier import DrawEfficientFrontier

from scientific.finance import expected_shortfall
from scientific.finance.efficient_frontier import SimulateEfficientFrontier

test plotly candle stick¶

In [2]:
with open('sp500.pickle', 'rb') as f:
    sp500 = pickle.load(f)
    
sp500.head(3)
Out[2]:
date open high low close volume
0 2014-09-17 1999.300049 2010.739990 1993.290039 2001.569946 3209420000
1 2014-09-18 2003.069946 2012.339966 2003.069946 2011.359985 3235340000
2 2014-09-19 2012.739990 2019.260010 2006.589966 2010.400024 4880220000
In [3]:
draw_candle_plotly(sp500[:20], 'Giá SP500')

test plotly line¶

In [4]:
draw_line_plotly(sp500[:20], 'date', 'close', 'Giá SP500')

test LogY Graph¶

In [5]:
with open('sp500_btc.pickle', 'rb') as f:
    sp500_btc = pickle.load(f)
    
sp500_btc.head(3)
Out[5]:
date open_sp500 high_sp500 low_sp500 close_sp500 volume_sp500 open_btc high_btc low_btc close_btc volume_btc
0 2014-09-17 1999.300049 2010.739990 1993.290039 2001.569946 3209420000 465.864014 468.174011 452.421997 457.334015 21056800.0
1 2014-09-18 2003.069946 2012.339966 2003.069946 2011.359985 3235340000 456.859985 456.859985 413.104004 424.440002 34483200.0
2 2014-09-19 2012.739990 2019.260010 2006.589966 2010.400024 4880220000 424.102997 427.834991 384.532013 394.795990 37919700.0
In [6]:
logy = LogYGraph((SeabornFn('lineplot', {'data': sp500_btc, 'x': sp500_btc.date, 'y': 'volume_sp500', 'color': 'red', 'label': 'S&P 500'}),
                SeabornFn('lineplot', {'data': sp500_btc, 'x': sp500_btc.date, 'y': 'volume_btc', 'color': 'green', 'label': 'Bitcoin'}),
                ))

fig, ax1, ax2 = logy.draw()
fig.suptitle('Trading volume of Bitcoine vs S&P500')
ax1.set_title('Linear scale')
ax2.set_title('Log scale')
ax1.grid(axis='y')
ax2.grid(axis='y')

Dist_Moments_Report¶

In [7]:
with open('dcp_df.pickle', 'rb') as f:
    dcp_df = pickle.load(f)
    
dcp_df.head(3)
Out[7]:
sp500_dcp btc_dcp gold_dcp
0 0.000000 0.000000 0.000000
1 0.023201 -0.125659 -0.038537
2 0.024534 0.112692 0.020424
In [8]:
Dist_Moments_Report(dcp_df)(95)
Out[8]:
⬇ means the lower the better, ⬆ otherwise
  sp500_dcp btc_dcp gold_dcp
volatility ⬇ 0.041 0.236 0.036
kurtosis ⬇ 1.642 0.069 0.305
VaR (at 95) ⬆ -0.065 -0.282 -0.053
CVaR (at 95) ⬆ -0.087 -0.330 -0.066
MDD ⬆ -0.200 -0.764 -0.161

Test CompareVarExtreme¶

In [9]:
CompareExtremeVaR(dcp_df, expected_shortfall)()
Out[9]:
  sp500_dcp btc_dcp gold_dcp
CVaR at 80 -0.048 -0.204 -0.042
CVaR at 90 -0.071 -0.275 -0.055
CVaR at 95 -0.087 -0.330 -0.066
CVaR at 99 -0.125 -0.382 -0.073
PCT - 80 -0.000 -0.000 -0.000
PCT - 90 0.474 0.351 0.318
PCT - 95 0.234 0.200 0.194
PCT - 99 0.434 0.156 0.102

Test Risk_Return_Ratio_Report¶

In [10]:
Risk_Return_Ratio_Report(dcp_df[['sp500_dcp', 'btc_dcp', 'gold_dcp']], dcp_df.sp500_dcp)()
Out[10]:
The higher the ratios the better
  sp500_dcp btc_dcp gold_dcp
Sharpe Ratio 0.264 0.361 0.139
Sortino Ratio 0.338 0.842 0.250
Information Ratio nan 0.328 -0.100
M2 Ratio 0.011 0.015 0.006

Test draw ECDF & ECDF Boundaries¶

In [11]:
with open('inflation_month.pickle', 'rb') as f:
    inflation_month = pickle.load(f)
    
inflation_month.head(3)
Out[11]:
close_sp500 close_btc gold_usd cpi_us inflation sp_pct btc_pct gold_pct
date
2014-09-30 1972.290039 386.944000 1223.565 238.031 0.000000 0.000000 0.000000 0.000000
2014-10-31 2018.050049 338.321014 1176.413 237.433 -0.251228 2.320146 -12.565897 -3.853657
2014-11-30 2067.560059 376.446991 1200.440 236.151 -0.539942 2.453359 11.269172 2.042395
In [12]:
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(20, 5), sharey=True)

ax1 = draw_ecdf(ax1, inflation_month.inflation, 'Inflation')
ax2 = draw_ecdf(ax2, inflation_month.btc_pct, 'Bitcoin')
draw_percentile_boundaries(inflation_month.inflation, ax1, 80, 99)
draw_percentile_boundaries(inflation_month.btc_pct, ax2, 80, 99)

Test percentile_comparision¶

In [13]:
with open('sample_percentile.pickle', 'rb') as f:
    sample_percentile = pickle.load(f)
    
for index, value in sample_percentile.items():
    print(index, value.head(3), sep='\n')

extreme
date
2015-03-31   -1.739611
2015-05-31    1.049138
2016-04-30    0.269940
Name: sp_pct, dtype: float64
middle
date
2014-09-30    0.000000
2015-02-28    5.489251
2015-04-30    0.852082
Name: sp_pct, dtype: float64
lowest
date
2014-10-31    2.320146
2014-11-30    2.453359
2014-12-31   -0.418859
Name: sp_pct, dtype: float64
In [14]:
percentile_comparision_data = [sample_percentile['extreme'], sample_percentile['middle'], sample_percentile['lowest']]
fig, ax = draw_boxplot_percentile_comparision(percentile_comparision_data, ['High CPI (top 10%)','Medium CPI','Low CPI (bottom 10%)'])
plt.title('Monthly S&P500 returns by inflation percentiles')
ax.set_ylabel('SP500 returns')
print('Triangle is Mean, Line is Median \n')
plt.show()

Triangle is Mean, Line is Median

Test Two Var Portfolio¶

In [15]:
with open('portfolio_data.pickle', 'rb') as f:
    portfolio_data = pickle.load(f)
    
portfolio_data.head(3)
Out[15]:
close_btc close_sp500 gold_usd
date
2014-09-30 386.944000 1972.290039 1223.565
2014-10-31 338.321014 2018.050049 1176.413
2014-11-30 376.446991 2067.560059 1200.440
In [16]:
bitcoin_ratio = (.2, .4, .6, .8)
sp500_ratio = (.2, .4, .6, .8)

tt = TwoVariablePortfolioReport(portfolio_data) 
tt.add_steps(bitcoin_ratio, sp500_ratio)
tt.create_portfolio_stat()
In [17]:
display(tt.create_volatility_rp('Bitcoin', 'SP500'), 
      tt.create_ratio_rp(sortino_ratio, 'bitcoin', 'sp500', 'Sortino'),
      tt.create_var_rp(value_at_risk_single, 90, 'bitcoin', 'sp500', 'VaR'),
      )
Volitality report
  20% Bitcoin 40% Bitcoin 60% Bitcoin 80% Bitcoin
20% SP500 0.053881 0.097706 0.144056 0.191102
40% SP500 0.055376 0.100339 0.147077 nan
60% SP500 0.059181 0.104221 nan nan
80% SP500 0.064891 nan nan nan
Sortino report
  20% bitcoin 40% bitcoin 60% bitcoin 80% bitcoin
20% sp500 1.096583 0.954729 0.900577 0.854240
40% sp500 1.024867 0.968305 0.859560 nan
60% sp500 0.877333 0.881725 nan nan
80% sp500 0.730776 nan nan nan
VaR at 90 percentile report
  20% bitcoin 40% bitcoin 60% bitcoin 80% bitcoin
20% sp500 -0.043533 -0.067770 -0.102013 -0.139887
40% sp500 -0.043704 -0.066070 -0.102028 nan
60% sp500 -0.049801 -0.078533 nan nan
80% sp500 -0.060850 nan nan nan

Test Efficient Frontier¶

In [18]:
efficient = SimulateEfficientFrontier(portfolio_data, 100)()


DE = DrawEfficientFrontier(efficient, 'volatility', 'mean', 'sortino ratio', 'index_col')
DE('Annualised Risk (Volatility)', 'Annualised Return').show()